Studies using dobutamine thallium-201 myocardial perfusion imaging have suggested a high sensitivity and specificity for the detection of coronary artery disease. However, few data are available comparing dobutamine with exercise stress for the detection and localization of perfusion defects. This study compared the effects of dobutamine and exercise stress using technetium-99m sestamibi single-photon emission computed tomographic imaging in the same patients in a prospective crossover trial. Twenty-four patients with a high likelihood of coronary artery disease underwent tomographic myocardial imaging at rest, after symptom-limited treadmill exercise, and after intravenous dobutamine (maximum 30 micrograms/kg/min). Tomograms of the left ventricle were divided into 20 segments and were interpreted without knowledge of patient identity or stress protocol. Dobutamine was well tolerated by all patients. Segment-by-segment concordance between exercise and dobutamine images was highly significant (kappa = 0.56, p < 0.0001). Global first-order agreement (normal vs abnormal) between exercise and dobutamine studies was 96% (kappa = 0.65, p = 0.02); global second-order agreement (normal vs fixed vs ischemic defect) was 88% (kappa = 0.45, p = 0.02). Regional first- and second-order agreement were 96 and 93%, respectively (p < 0.001 for both). Twenty patients underwent coronary angiography. Comparisons between exercise and angiography and between dobutamine and angiography were similar for both global agreement (95 vs 100%, p = NS) and regional agreement (77 vs 72%, p = NS).(ABSTRACT TRUNCATED AT 250 WORDS)